Anthropic report alleges ~200M distillation attacks on Claude by Alibaba, Moonshot AI and DeepSeek
scaling01 · x · 2026-09-11
A new Anthropic report alleges persistent and escalating distillation attacks by China-based AI labs to harvest Claude's most valuable capabilities, including agentic tool use, coding, and reasoning. Per TechCrunch: Alibaba logged 151M+ exchanges from May–July 2026 peaking near 3M/day, Moonshot AI 23M+, and DeepSeek 12.1M+ in 14 days — nearly 200M exchanges total. Anthropic says unauthorized labs have developed increasingly sophisticated methods to circumvent its defenses. It previously called out specific labs in February, and OpenAI has reported similar activity attributed to DeepSeek. Note this is Anthropic's own report; the named companies have not yet responded.
More from Models
- GPT-6 Astra drives a robot arm on first try via physical ICL; Ken Goldberg touts Agentic Robotics — zhaoran_wang · 2026-09-11
- OpenAI's internal model claims a Navier-Stokes millennium prize proof, says analyst — QuintinPope5 · 2026-09-11
- DeepSeek 4.1 flash reportedly uses large ngram embeddings, echoing Qwen4 architecture — ccerrato147 · 2026-09-11
- ValsAI launches RSI Index, first third-party benchmark measuring how close AI is to self-improvement — JenniferHli · 2026-09-11
- Assistant Benchmark goes live: 61 assistants scored across 15 real-use dimensions — Scobleizer · 2026-09-11
- OpenAI pauses new $200 ChatGPT Pro signups as GPT-6 Astra demand overwhelms capacity — 机器之心 · 2026-09-11